Will online analysis (w/decennial Census & ACS) give estim for fams w/related child(ren) or have to use extract?

#1

When I try to use the online analysis tool with ACS PUMS or Census 2000 PUMS data, I find that there are no household level or family level variables for related child. If I try to use the relate person-level variable, I find there is no way to apply a household weight that will give me an estimate of families in which there is a related child(ren). Do I have to use get an extract of the data instead? In any case, could you please explain how I should weight the data to provide the desired results.

I’m using 2009-2011 ACS PUMS. I want to get estimates of the numbers and percentages of families with related child(ren) under 18 years of age (compared to the number of all other households) who fall into different income categories based on income to poverty ratios. (By related children, I mean children related to the householder.)

We figured out a way to do this analysis online for families with own child(ren) < 18 , but not for families with related child(ren) < 18.

#2

For this specific question it may be easier to use an extract as it will allow you to use more sophisticated methods for generating a subset of the data. Since you are interested in counting families, I would recommend looking at the FAMUNIT variables, used in conjunction with SERIAL. However, there is no weight that directly corresponds to FAMUNIT. Some researchers choose to select one individual per family and use that individual’s PERWT as the family weight.